# VoxCPM2 Technical Report

arXiv: https://arxiv.org/abs/2606.06928 · HTML: https://mlsanity.com/paper/2606.06928/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Jun 2026.

## Built on by (most cited first, 17 of 17)

- [Aslema at NADI 2026: Data Augmentation for Intent Recognition and Slot Filling](https://mlsanity.com/paper/2608.18689/) (2026)
- [AuK Technical Report: An Open-Source Foundational Model for Speech Generation and Editing](https://mlsanity.com/paper/2609.08936/) (2026)
- [Closing the Quality Gap in Low-Resource Text-to-Speech: LoRA Fine-Tuning of VoxCPM2 for Khmer and Korean](https://mlsanity.com/paper/2606.26618/) (2026)
- [Fr'echet Distance Loss on Speech Representations for Text-to-Speech Synthesis](https://mlsanity.com/paper/2607.06027/) (2026)
- [BlueMagpie-TTS: A Token-Efficient Tokenizer, Language Model, and TTS for Taiwanese-Accent Code-Switching Speech](https://mlsanity.com/paper/2607.06054/) (2026)
- [FreyaTTS: A Compact Tokenizer-Free Flow-Matching Transformer for Turkish-First Speech Synthesis](https://mlsanity.com/paper/2607.09530/) (2026)
- [Stable Autoregressive Speech Generation with Low-Frame-Rate High-Dimensional Continuous Tokens](https://mlsanity.com/paper/2607.29363/) (2026)
- [CuteTTS: Efficient and High-Quality Speech Synthesis via Autoregressive Modeling of Continuous Latents](https://mlsanity.com/paper/2608.08638/) (2026)
- [Confucius4-TTS: Transcript-Free Cross-Lingual Zero-Shot TTS with a Learnable Speaker Encoder](https://mlsanity.com/paper/2608.11650/) (2026)
- [FastThaiG2P: Lightning-fast Thai Grapheme-to-phoneme Conversion for Voice Agent Pipelines](https://mlsanity.com/paper/2608.12814/) (2026)
- [HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice Grounding](https://mlsanity.com/paper/2608.29120/) (2026)
- [COT-TTS: Audio Context-Aware Text-to-Speech with Chain-of-Thought Reasoning](https://mlsanity.com/paper/2609.22697/) (2026)
- [Learnable Classifier-Free Guidance Null Embeddings for Enhanced Controllable Speech Synthesis](https://mlsanity.com/paper/2609.25411/) (2026)
- [NVAlign: Direct-Gradient Optimization for Non-Verbal Control in Continuous Autoregressive Flow Matching Text-to-Speech](https://mlsanity.com/paper/2609.31892/) (2026)
- [A Comprehensive Objective Evaluation of Modern Text-to-Speech for Turkish Using Speech Quality Assessment Models](https://mlsanity.com/paper/2610.06057/) (2026)
- [VoTSpeech: Decoupling Voice Design from Speech Generation in Instruction-Following TTS](https://mlsanity.com/paper/2610.14694/) (2026)
- [Training a Compact, Strong, and Realistic Text-to-Speech Model](https://mlsanity.com/paper/2610.14776/) (2026)

## Abstract (arXiv)

We present VoxCPM2, a https://info.arxiv.org/help/prep#abstractsfully open-source multilingual and controllable speech generation foundation model that extends the hierarchical diffusion-autoregressive modeling paradigm of VoxCPM. VoxCPM2 advances the framework in three key dimensions: (i) capability, by unifying 30 languages, 9 Chinese dialects, natural-language voice design, style-controllable voice cloning, and high-fidelity continuation cloning within a single backbone; (ii) quality, through an asymmetric AudioVAE that encodes at 16 kHz and reconstructs at 48 kHz, enabling implicit super-resolution with high encoding efficiency; and (iii) scale, by jointly scaling the model to 2B parameters and the training data to over 2 million hours of multilingual speech. To support these diverse capabilities within one model, we introduce a unified sequence organization that expresses all generation modes through different arrangements of the same input building blocks, allowing joint training under a single set of parameters and objective. VoxCPM2 achieves state-of-the-art or competitive performance on public zero-shot and instruction-following TTS benchmarks. On our internal 30-language evaluation set, it attains an average WER of 1.68%. These results demonstrate that hierarchical continuous-latent modeling, without relying on any external discrete speech tokenizer, offers a viable and powerful foundation for large-scale multilingual and controllable speech generation. The model weights, fine-tuning code, and inference tools are publicly released under the Apache 2.0 license to foster community research and development.